tech
Anthropic says it will watermark text generated by its AI models
Anthropic will extend support for watermarking AI generations for older models as well.
TL;DR
- Anthropic will watermark text generated by its AI models, including Claude, to comply with EU regulations.
- The EU AI Act's Transparency Code requires AI-generated or edited content to be marked for identification.
- Watermarking will be applied at the model level and will persist when text is copied and pasted.
- Anthropic will extend support for watermarking to older models and is using the C2PA open standard for files.
- Other companies like Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia are also committed to the EU's code.
- Platforms like Suno and Substack are also implementing measures to mark AI-generated content.